MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I call a set PECULIAR, if its elements are uncountable, pairwise disjoint subsets of R (the real number system). As for example, the set {(0,1),(3,5),[8,9]\Q},where Q denotes the set of rationals, is peculiar. My first question may have a very trivial answer that I cannot see immediately,that does there exist an uncountable, peculiar set? Next, I define the STRENGTH of a peculiar set to be the union of all its elements (which are sets). My question is that what are the subsets of R (if any), that are the strength of some uncountable peculiar set? In case complete specification seems impossible or meaningless, some examples will do.

share|cite|improve this question
i do not think that this question is appropriate for math overflow. Every uncountable set is the disjoint union of uncountably many uncountable sets. – Goldstern Dec 17 '11 at 16:34
Take a bijection $f\colon\mathbb R\to\mathbb R\times\mathbb R$ and define the peculiar collection $\{\{f^{-1}(a,x)\mid a\in\mathbb R\}\mid x\in\mathbb R\}$. – Asaf Karagila Dec 17 '11 at 16:41
up vote 9 down vote accepted

Since the OP asked about sets of reals, and since these cannot be amorphous (because no amorphous set admits a linear ordering), let me point out that it is consistent with ZF that there is a Dedekind finite (hence not countable, as in Joel's answer) set of reals that cannot be partitioned into uncountably many many infinite pieces. One model in which this happens is the basic Cohen model, as defined in Section 5.3 of Jech's book "The Axiom of Choice". The relevant Dedekind finite set is the set of Cohen reals explicitly added by the forcing (the set Jech calls $A$). The fact that it cannot be split into uncountably many uncountable sets follows from work of Halpern and Levy ("The Boolean prime ideal theorem does not imply the axiom of choice," Axiomatic Set Theory, Part I, Proc. Symp. Pure Math. XIII, Part I (1971) pp. 83-134).

share|cite|improve this answer
This set $A$ cannot even be split into infinitely many infinite sets, regardless to uncountability of either the partition or the parts. – Asaf Karagila Dec 17 '11 at 19:41
@Asaf: If you view the elements of $A$ as subsets of $\omega$, won't there be infinitely many whose first element is $n$, for each $n\in\omega$? – Andreas Blass Dec 17 '11 at 22:25
@Andreas: While it does sound as a very convincing argument, I have run into several proofs that this set is a counterexample to the assertion that every set has a group operation, by the fact that every partition has only finitely many non-singletons. At this moment I am too tired to verify all the details of either arguments, and it will have to wait for tomorrow. – Asaf Karagila Dec 17 '11 at 22:38
See Justin's answer in this link:… In Cohen's basic model the D-finite set of reals is "strong" in the amorphous sense of the word (no partition into infinitely many non-singletons exists). – Asaf Karagila Dec 18 '11 at 23:05
@Asaf: Justin wrote that $A$ can't be partitioned into finite sets, and that proof looks OK to me. But I don't see how that leads to your claim that $A$ can't be partitioned into infinitely many infinite sets. Specifically, if I try to carry out a proof like Justin's in the case where the pieces are allowed to be infinite, there's no condition like Justin's $q$ that decides which $a_i$'s (with a ground-model set of indices $i$) constitute a particular piece of the partition. – Andreas Blass Dec 19 '11 at 0:39

The comments have answered the question in the context of ZFC, where the axiom of choice is available.

But there is something interesting to say when the axiom of choice fails (and this answer was too long for a comment). Namely, it is consistent with ZF that not every uncountable set can be partitioned into uncountably many disjoint uncountable sets. Indeed, it is consistent with ZF that an uncountable set cannot even be partitioned into two disjoint infinite sets. This is because it is consistent with ZF that there is an amorphous set, an infinite set all of whose subsets are either finite or the complement of a finite set. It follows that amorphous sets cannot be countable, since if there were an enumeration we could produce bad subsets by taking every other element. Thus, amorphous sets are uncountable sets with no partition into two or more disjoint infinite sets.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.